The game of roulette requires that each roulette wheel is well balanced and without defects in order to generate an even distribution of probabilities for all the numbers of the roulette wheel. This is important both from the perspective of the player and the casino, as the winning number of the roulette wheel will become affected. Further, the affected features of a particular roulette wheel may be exploited by an observant player, i.e. a certain outcome of a round may be more frequently appearing than the theoretical probability, leading to a potential loss to the casino.
It is thus important to continuously monitor and check each roulette wheel of a casino, in order to detect any tendencies to defects or bias at an early stage.
Such defect or bias may be the result of dust, dirt and/or body exudate emitted by the croupier, errors of the roulette wheel due to mistakes during the wheel manufacturing process, etc. Also the table on which the roulette wheel is placed may be uneven or non-horizontal. It may also be the result of manipulation or inadvertent placing of the roulette wheel.
A possible method to track a roulette wheel bias may be to continuously keep track of the historical outcome of numbers of the roulette wheel. A problem with this method is that it is reactive, not proactive; i.e. it is only possible to afterwards detect that the roulette wheel has had a bias so serious that the outcome has been affected. Thus such statistical methods based on winning number analysis alone is insufficient. It would be desired to spot any tendencies of deviation from an expected outcome of the roulette wheel at a very early stage, preferably in a proactive manner, before the winning number statistics of the roulette wheel is affected.
Document U.S. Pat. No. 8,028,993 B2 describes a roulette wheel comprising three ball detectors/sensors for detecting a ball and determining where the ball drops into the rotor of the roulette wheel. The sensors are in-built in the roulette wheel. Thereby, a drop zone bias may be detected by storing the detected respective zones where the ball drops into the rotor. A first problem with this approach is that the number of sensors is limited to three sensors; thereby also the number of drop zones becomes limited, which is insufficient for an accurate drop zone determination. A second problem is that only particular roulette wheels having the ball detectors in-built could be used. It would be desired to instead discover a general method for detecting any roulette wheel bias which is applicable on any arbitrary roulette wheel.
Document US20140098219A1 describes a method for identifying defects in a roulette wheel. A first trajectory of a roulette ball may be determined after launch of the roulette ball by capturing movement of the roulette ball on the roulette wheel. The determining step may be repeated to determine additional trajectories, and a plurality of areas that the roulette ball avoided during travel along the trajectories may be identified. A graphical representation of the plurality of avoided areas may be generated to identify regions of the roulette wheel that include defects.
Thereby, much data has to be stored, which consumes processing resources and also delay the processing.
Document EP2814010A1 relates to a method for immersive roulette gaming. The methods comprise receiving video data from a plurality of image sensors; detecting a ball on a roulette wheel based on the video data using a hardware processor; obtaining motion data about the ball based on the video data using the hardware processor; determining whether the ball is about to fall into a pocket of the roulette wheel based on the video data using the hardware processor; and producing multiple slow-motion images of the ball in response to determining that the ball is about to fall into a pocket on the roulette wheel.
The described solution does not concern detection of a roulette wheel bias at all but rather a solution for providing immersive roulette gaming.
It is thus desired to improve roulette wheel bias tracking.